KR102258722B1 - Dynamic Blockchain System and Operating Method thereof - Google Patents

Abstract

The present invention relates to a dynamic blockchain system and a method of operation thereof. The dynamic blockchain system according to the present invention includes a plurality of terminal nodes owned by a user, and a transaction node to which a transaction is transmitted is mapped to each terminal node. Terminal node pool; And a plurality of transaction nodes that asynchronously process a transaction transmitted from each terminal node of the terminal node pool and perform consensus of a virtual block by mapping the next transaction node to each transaction node, wherein the plurality of transaction nodes Network performance can be improved by providing a pool of transaction nodes that automatically balance the loads of users.

Description

Dynamic Blockchain System and Operating Method thereof

The present invention relates to a dynamic blockchain system and an operation method thereof, and to a blockchain system capable of dynamically load balancing and an operation method thereof.

Blockchain technology, which started with Bitcoin, is attracting attention from numerous industries as an innovative technology that will lead the next-generation Internet, and rapidly with the advent of Ethereum, Hyperledger, Ripple, and R3. It is expanding its influence industrially. In addition, competition for technologies and platforms is intensifying with the birth of a startup that provides blockchain solutions in various fields such as virtual currency, asset management, sharing economy, IoT, health, and logistics.

Blockchain system is a system in which numerous nodes are connected via a P2P network to process user transactions, and it can be seen as a kind of distributed ledger system that sequentially stores records of transactions. Once recorded, it is almost impossible to change. It has a characteristic. The blockchain system must ensure that all nodes have the processing record for the same transaction, and the consensus algorithm makes it possible.

Blockchain systems can be largely divided into permissionless blockchains and permissioned blockchains depending on whether or not nodes participating in the blockchain are restricted from participating. In the case of a permission-type blockchain, a completely different form of consensus algorithm can be used than the proof-of-work-based consensus algorithm used in an unlicensed blockchain. This is because participating nodes are limited, so an algorithm that can take advantage of that point can be applied. A representative algorithm is PBFT (Practical Byzantine Fault Tolerant).

Hashgraph algorithm is a consensus algorithm that can be used for permissioned blockchain. Hashgraph is an algorithm that works by passing gossip to other unspecified nodes rather than a linked list method to overcome the limitations of the existing blockchain. While maintaining the anti-forgery function and decentralized decision-making structure, which are the advantages of the existing blockchain, it can process hundreds of thousands of transactions per second.

This hash graph is fair because a specific user cannot change the order of transactions. For example in the stock market, suppose that Ellis and Bob simultaneously buy the last one left at the same price. In the blockchain system, if a miner puts both transactions in the same block, the order can be arbitrarily determined, or only Ellis' transaction and Bob's transaction can be included in the next block. In the hashgraph, in this case, an individual cannot change the order of transactions. Here, the best way for Elise to spread to all users in the distributed system before Bob's transaction is to invest in internet speed, thereby promoting fair competition. Since the transaction processing authority has not been delegated to a specific person, Ellis is unable to obtain unfair benefits by bribing miners.

In addition, in the hash graph, it is fair because a specific user cannot stop or stop a transaction entered into the system. In the blockchain system, if multiple miners refuse to include a specific transaction in a block, processing may be delayed until one or two mining periods. However, in the hashgraph, you cannot block users' transactions except by blocking the Internet connection.

However, in the hash graph algorithm, if the load continues to be concentrated on one node, a bottleneck may occur even when real-time payment processing is required, and the overall processing may be delayed, and flexible processing is difficult because the number of nodes in the network is fixed. In addition, events stored in the physical node by the hash graph algorithm are accumulated in the volatile memory and are vulnerable to external attacks. In addition, data accumulated in the memory may be erased when the power is cut off.

In order to solve the above problems, an object of the present invention is to provide a dynamic blockchain system and a method of operation thereof.

An object of the present invention is to provide a block chain system capable of load balancing and an operation method thereof.

In addition, an object of the present invention is to provide a block chain system and a method of operation thereof that can prevent damage from external attacks and protect data.

In order to achieve the above object, a dynamic blockchain system according to an embodiment of the present invention includes a plurality of terminal nodes owned by a user, and a transaction node to which a transaction is to be transmitted is mapped to each terminal node. pool; And a plurality of transaction nodes that asynchronously process a transaction transmitted from each terminal node of the terminal node pool and perform consensus of a virtual block by mapping the next transaction node to each transaction node, wherein the plurality of transaction nodes It is possible to provide a pool of transaction nodes that automatically perform load balancing of users.

In the dynamic blockchain system, the transaction node pool may further include a transaction management node that automatically performs load balancing of the plurality of transaction nodes, and the transaction management node transmits each time a virtual block is agreed upon for a transaction. It is possible to automatically perform load balancing of the plurality of transaction nodes by determining the load state of the specific transaction node from the notification information of the specific transaction node.

The transaction management node may asynchronously process a transaction transmitted from each terminal node of the terminal node pool and map a predetermined number of transaction nodes to one channel in order to perform virtual block consensus.

The transaction management node may determine a load state of the specific transaction node of a specific channel and, if the load is greater than or equal to a threshold, map a predetermined number of transaction nodes to a new channel and branch from the specific channel.

The dynamic blockchain system may further include a block node pool for generating a block chain by collecting at least hundreds of transactions transmitted after consensus of virtual blocks for each transaction in the transaction node pool to create a block chain.

The block chain transaction created from the block node pool may include a transmission order of the predetermined number of transaction nodes related to the corresponding transaction agreement.

The terminal node pool may further include a terminal management node for mapping the plurality of terminal nodes and a transaction node to which a transaction is to be transmitted from each terminal node, and the terminal management node manages not only the mapping but also membership registration and authentication. It can perform a server function.

According to another embodiment of the present invention, a method of operating a dynamic blockchain system including a terminal node pool including a plurality of terminal nodes owned by a user and a transaction node pool including a plurality of transaction nodes is provided in each terminal node. Mapping a transaction node to which a transaction is to be transmitted, and a next transaction node to each transaction node; Transmitting a transaction from one of the plurality of terminal nodes to a mapped transaction node; Asynchronously processing a transaction transmitted from the terminal node in the plurality of transaction nodes and performing virtual block consensus; And automatically performing load balancing of the plurality of transaction nodes.

With the above-described configuration, the present invention can dynamically increase the performance of the network by branching or re-merging according to the load state in the node.

In the present invention, it is also possible to divide physical nodes and logical nodes to uniformize the availability of a network through mapping and increase defense power against external attacks.

The present invention can also prevent damage from external attacks and protect data by creating and managing a block chain from a pool of block nodes.

1 is a block diagram of a dynamic blockchain system according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of a data stack usable in the block chain system of FIG. 1.
3 is a diagram illustrating an example of a block structure stored in a block node.
4 is a diagram illustrating an example of a general hash graph gossip protocol.
5 is a diagram showing an example of branching of a hash graph according to the present invention.
6 is a diagram showing an example of merging of hash graphs according to the present invention.
7 is a diagram illustrating an example of internal processing of channels generally allocated to transaction nodes.
8 is a diagram showing an example of internal processing of a channel branch according to the present invention.
9 is a diagram showing an example of internal processing of channel merging according to the present invention.
10 is a diagram illustrating a flowchart of a method of operating a dynamic blockchain system according to an embodiment of the present invention.

Hereinafter, a preferred embodiment of a dynamic blockchain system and a method of operation thereof according to the present invention will be described with reference to the accompanying drawings. For reference, in the following description of the present invention, the terms referring to the components of the present invention are named in consideration of the functions of the respective components, and thus should not be understood as limiting the technical components of the present invention. Will be.

1, the dynamic blockchain system 100 includes a terminal node pool 110, a transaction node pool 120, and a block node pool 130.

The terminal node pool 110 is related to a terminal owned by a user, for example, a plurality of terminal nodes 112 for requesting actual payment and a terminal management node 114 for load balancing of the terminal nodes 112 ) Can be included.

The plurality of terminal nodes 112 may be configured with #01 to #40, for example, as shown in FIG. 1. In addition, the terminal management node 114 may appropriately distribute the transaction nodes mapped to each terminal node 112 so that the load is appropriately distributed among the plurality of terminal nodes 112 to a small number of transaction nodes 122. The terminal management node 114 may perform load balancing as well as monitoring to detect abnormal processing and external intrusion. When an abnormal processing or external intrusion is detected, the terminal management node 114 may respond according to a previously set security policy. The terminal management node 114 may appropriately allocate the transaction nodes 122 to be transmitted to each terminal node 112, as shown in FIG. 1.

In the terminal management node 114, a server that manages membership registration and authentication may play this role, and a separate enhanced security solution may be mounted. In general, before making payment, it is essential to authenticate whether the user really owns this terminal, and the terminal management node 114 checks the result using authentication means such as biometric authentication, password, or pattern at the terminal node. Thus, abnormal access can be detected.

The transaction node pool 120 may include a plurality of transaction nodes 122 for processing transactions and a transaction management node 124 for load balancing of the transaction nodes 122. The transaction management node 124 bundles and manages a predetermined number of transaction nodes 122 into a single channel, and when a large load is applied to one channel, a new channel can be appropriately branched to allocate a new channel. Conversely, too little load is required. If nodes are wasted, these channels can be consolidated into a single channel. In addition, the transaction management node 124 may designate and transmit the next transaction node 124 of each transaction node 122 or may randomly designate and transmit the next transaction node 124. The transaction management node 124 may also perform monitoring to detect abnormal processing and external intrusion, and when abnormal processing or external intrusion is detected, may respond according to a previously set security policy.

As shown in FIG. 1, the plurality of transaction nodes 122 may be configured from &01 to &20, for example. In FIG. 1, five transaction nodes 122 are shown as one channel. The order of transmission in the transaction node pool 120 shown in FIG. 1 is determined for convenience of description, and may be in the form of a hash graph that is a consensus algorithm according to a specific embodiment.

The block node pool 130 includes a plurality of block nodes 132 that generate a block including a plurality of transactions identified in one channel, and a block management node 134 that manages a channel of nodes for block generation. can do. The plurality of block nodes 132 may be composed of, for example, %01 to %04, as shown in FIG. 1. The block management node 134, although not shown in FIG. 1, may also distribute or consolidate the load like the transaction node pool 120 to manage.

The above management nodes are basically redundant to synchronize data regularly. If the main management node operates abnormally, the sub-management node immediately acts as the main management node, and the main management mode can enter the recovery mode. .

1, in the embodiment of the present invention, the transaction node 122 and the terminal node 112 are clearly separated. However, in the transaction node 122 and the block node 132, one node may perform both functions depending on implementation.

Although bidirectional communication is allowed between the management nodes shown in FIG. 1, it is preferable that normal nodes are configured to receive only instructions from the management node. That is, when an event of configuration change such as a change of a mapping node occurs, general nodes operate according to the order of commands of the management node, and do not feed back to the management node or transmit separate data. Accordingly, it is possible to fundamentally prevent a hacker from invading the management node through a general node.

Also, it usually becomes difficult to block if a hacker has already infiltrated the network. Therefore, the security of the terminal management node needs to be strengthened so that the terminal management node corresponding to the network gate detects an abnormal intrusion and is sufficiently resistant to attack.

FIG. 2 is a diagram illustrating an example of a data stack usable in the block chain system of FIG. 1.

In general, payment can be divided into a process of sending money and a process of receiving money. In FIG. 2, a data stack is shown as an example of a process of transmitting money during a payment process.

When the user of the terminal node #01 wants to send money to another terminal node, the terminal node #01 includes the terminal ID information along with specialized data such as "Gap sends coin 1 to Eul", for example. The transaction node (Tr Node) id is transmitted as one transaction. Here, the terminal node ID may be recorded as its own terminal node #01 as a source, and the transaction node ID may be recorded as a destination, and the transaction node &01 designated by the terminal management node 114 may be recorded.

In addition to the transaction information received from terminal node #01, transaction node &01 adds and transmits a next transaction node, for example, transaction node &02. In addition to the information received from transaction node &01, transaction node &02 adds and transmits a next transaction node, for example, transaction node &03. In addition to the information received from transaction node &02, transaction node &03 adds and transmits a next transaction node, for example, transaction node &04. In addition to the information received from transaction node &03, transaction node &04 adds and transmits a next transaction node, for example, transaction node &05.

In addition, when consensus of a virtual block is reached according to a gossip protocol such as a hash graph algorithm, transaction node &05 adds and transmits a block node, such as Bl Node %01, in addition to the information received from transaction node &04.

In the case of a hash graph in transaction nodes, a transaction becomes an event, and the events are stored in association with each other. However, the data stored in the transaction nodes are not stored in the blockchain but are events stored in volatile memory, so they are not only vulnerable to external attacks, but also the data accumulated in the memory may be erased when the power is cut off. In addition, in the transaction node pool 120, it is not easy to execute a smart contract for signing and executing various types of contracts.

Block nodes not only enable smart contracts, but also manage data with a block chain, so you can take advantage of the advantages of a block chain. Block node %01 can generate one block when, for example, 2000 transactions with consensus of virtual blocks are received.

3 is a diagram illustrating an example of a block structure stored in a block node.

The block is divided into a header and a body, and in the header, the version number, block creation time, previous block hash value, consensus metadata, block metadata, and hash value of the current block can be recorded, and transactions can be recorded in the body. .

As described above, block n must include, for example, not only 2000 transactions, but also hash values of the previous block n-1 and block n in addition to the hash value of the transaction. Since the hash value of block n is recorded in the hash value area of the previous block of block n+1, the blocks have a connection relationship with each other. In addition, in the consensus metadata, the number of transaction nodes related to the consensus of the transaction may be recorded.

Hereinafter, processing in the transaction node pool 120 will be described in more detail.

As shown in FIG. 1, transactions, for example, events in the case of a hash graph, are received from a plurality of terminal nodes 112 and a plurality of transaction nodes 122 in the transaction node &01 122. In order to store these events, each transaction node 122 is provided with a Kafka cluster (not shown). Events stored in the Kafka cluster are processed by the corresponding transaction node 122 according to the input order.

4 is a diagram showing an example of a general hash graph gossip protocol, FIG. 5 is a diagram showing an example of a branch of a hash graph according to the present invention, and FIG. 6 is a diagram of a merge of hash graphs according to the present invention. It is a figure showing an example. Here, each node is a transaction node mapped to the terminals of Alice, Bob, Carol, Dave, and Ed, respectively.

4, too much load, that is, transactions are concentrated on the Bob node. In this case, processing of the Bob node is delayed due to a lot of load, and other transaction nodes belonging to the same channel are also affected, and the overall processing may be delayed. When it is determined that the overall processing speed of the same channel is slowed down due to the heavy load of the Bob node, the transaction management node 124 creates a Bob2 node, as shown in FIG. 5, when, for example, transactions are concentrated and exceed the threshold values. Afterwards, the Kafka cluster on the Bob node causes some of the transactions to be processed to be executed on the Bob2 node. On the other hand, branching in the channel of the hash algorithm may be properly handled by the transaction node.

On the other hand, referring to FIG. 4 again, the use of the Dave node and the Ed node is significantly reduced. In this case, since unnecessary power consumption occurs, the transaction management node 124 merges the Ed node into the Dave node and shares it as one transaction node, as shown in FIG. 6. In this case, the transaction management node 124 can change the next transaction node of the transaction node so that the transactions of the Ed node are transmitted to the Dave node, and also change the terminal management node 114 to change the destination from the terminal node to the transaction node. You can send a request signal. If you want to search for your own transaction details in the block created later, you can check all processing with the terminal id because the block contains both the transaction node created by your terminal id and the next transaction node in the transaction node.

7 is a diagram showing an example of internal processing of a channel generally allocated to transaction nodes, FIG. 8 is a diagram showing an example of internal processing of a channel branch according to the present invention, and FIG. 9 is a diagram showing an example of internal processing of a channel branch according to the present invention, and FIG. It is a figure which shows an example of the internal process of channel merging according to.

Transaction nodes concentrated in the physical space are vulnerable to security, and once the network is breached, there is a risk of damage to the entire network. Therefore, if nodes are distributed in several regions, it is advantageous to cope with external attacks and conform to the original purpose of the blockchain. Therefore, it is desirable to configure a distributed network using a mapping table. This makes it possible to have a network that is logically continuous, but is physically spread across multiple regions. In other words. If there are 20 nodes in the actual network and one channel uses 5 nodes, the logical node address is [&01, &02, &03. &04, &05], but the actual operating physical node address is [&12, &19, &03. &05, &10] can be randomly distributed.

7 to 9 are diagrams illustrating each of FIGS. 4 to 6 in detail in terms of channel mapping, channel branching, and channel merging.

As shown in FIG. 7, in the present invention, two logical nodes according to channels may be used for one physical node according to an embodiment. For example, the transaction management node 124 may allocate multiple channels to one physical node to balance the load.

If the maximum processing capacity of the physical node is 20Tr/sec, the processing capacity of channel A is 10Tr/sec, the processing capacity of channel B is 15Tr/sec, and the processing capacity of channel C is 10Tr/sec, In physical nodes, the transaction is processed because the throughput is delayed by 5Tr/sec, so the processing of the transaction in the channel is significantly delayed. Since the transaction management node 124 has a processing capacity of 10Tr/second in the following physical nodes, as shown in FIG.8, the transaction management node 124 creates a channel D by branching the channel B and processes a processing capacity of 5Tr/second. You can do it.

On the other hand, if the maximum processing capacity of the physical node is 20Tr/sec, the processing capacity of channel A is 5Tr/sec, the processing capacity of channel B is 5Tr/sec, and the processing capacity of channel C is 5Tr/sec, Since sufficient processing capacity to process channel C remains in the above physical nodes, channel C can be branched and merged into channel A, as shown in FIG. 9.

Table 1 below compares the characteristics of the dynamic blockchain system of the present invention with the published Hyperledger fabric and hash graph.

division Hyperledger Fabric Hash graph Dynamic Blockchain System Distribution of node functions There is no 0.6. Functional differentiation is possible from 1.0. none. There are terminal nodes, transaction nodes, and block nodes, and also each management node. Consensus algorithm BFT-Slow because the time line must be aligned at all nodes in order to proceed one step synchronously. ABFT-Fast because it can handle continuous events asynchronously. Dynamic ABFT-Faster than a hash graph because it can load balance asynchronously or automatically like a hash graph. Block creation Distributed Ledger (Distributed Ledger), for processing speed, wasted a lot of space due to a small transaction per block for processing speed, but 1.0 zoo-keeper was introduced to store it in the memory as a heap. Instead, it requires a lot of memory. Like Tangle, DAG (Directed Acyclic Graph) is present in memory, and there is no block. All data is lost when the system is down. Commited Transaction becomes Virtual Block and approval is completed. In addition, a separate block generation network distributes the block generation load. Low memory usage is possible and storage space efficiency is high as 2000 transactions make 1 block like Bitcoin. Smart Contract Run the chain code on the virtual machine. none. It is based on Hyperledger Fabric and is handled by the management node. Extensibility Sharding technique can be used by using multi-channel. Not yet materialized. Throughput can be increased by increasing the number of nodes and automatic load balancing is possible.

10 is a diagram illustrating a flowchart of a method of operating a dynamic blockchain system according to an embodiment of the present invention.

The transaction management node 124 creates one channel by connecting a predetermined number of transaction nodes 122 (S1002). The transaction management node 124 maps the next transaction node to which a transaction is to be transmitted for each transaction node 122 of each channel (S1004). The transaction management node 124 maps the block node 132 to which the transaction is to be transmitted to the last transaction node 122 to be transmitted in each channel (S1006). The terminal management node 114 maps the transaction node 122 to which a transaction is to be transmitted when the new terminal node 112 is subscribed (S1008).

Thereafter, a transaction is transmitted to one of the plurality of terminal nodes 112, for example, the transaction node 122 mapped in #01 (S1010). The plurality of transaction nodes 122 of the corresponding channel asynchronously process the transaction transmitted from the terminal node 112 and perform consensus of the virtual block (S1012).

The transaction management node 124 determines the load state of the specific transaction node 122 from the notification information of the specific transaction node 122 transmitted whenever a virtual block is agreed upon for one transaction (S1014).

The transaction management node 124 automatically performs load balancing of the plurality of transaction nodes 122 according to the load state of the specific transaction node 122 (S1016). For example, if the load of the specific transaction node 122 is greater than or equal to the first threshold, the transaction management node 124 may map a predetermined number of transaction nodes 122 to a new channel and branch from a specific channel. If is less than or equal to the second threshold, other channels may be merged into the corresponding channel.

The block node 132 creates a block chain by collecting at least hundreds of transactions transmitted after the consensus of the virtual block for each transaction and generating a block (S1018).

The embodiments of the present invention described above are merely illustrative of the technical idea of the present invention, and the protection scope of the present invention should be interpreted by the following claims. In addition, those of ordinary skill in the art to which the present invention pertains will be able to make various modifications and variations without departing from the essential characteristics of the present invention, and all technical ideas within the scope equivalent to the present invention are the present invention. It should be construed as being included in the scope of rights of

100: dynamic blockchain system 110: terminal node pool
112: terminal node 114: terminal management node
120: transaction node pool 122: transaction node
124: transaction management node 130: block node pool
132: block node 134: block management node

Claims (10)

A terminal node pool including a plurality of terminal nodes owned by a user, and in which a transaction node to which a transaction is to be transmitted is mapped to each terminal node;
A plurality of transaction nodes performing consensus of a virtual block-Each of the plurality of transaction nodes has a transaction node to be connected next to each transaction node, so that each terminal node of the terminal node pool performs a transmitted transaction. Asynchronous processing-and, a pool of transaction nodes including a transaction management node that automatically performs load balancing of the plurality of transaction nodes-a transaction node to be connected next is mapped to each of the plurality of transaction nodes, Under the control of the transaction management node, a transaction node in which a predetermined transaction is to be performed may be selected; And
In the transaction node pool, it includes a block node pool to create a block chain by collecting at least hundreds of transactions transmitted after consensus of virtual blocks for each transaction, and creating a block,

The transaction management node,
Map each transaction node to one channel,
Based on the notification information of the transaction node transmitted whenever a virtual block is agreed upon for a predetermined transaction, the load state of the transaction node of a specific channel is determined,
If the load of the transaction node is greater than or equal to a threshold, an additional transaction node is added to the specific channel,
A dynamic blockchain, characterized in that load balancing is automatically performed by dynamically distributing transactions transmitted to the transaction node of the specific channel to the transaction node of the specific channel and the additional transaction node of the specific channel. system. delete delete delete delete The method of claim 1,
A dynamic blockchain system, characterized in that the transaction of the block chain created from the block node pool includes a transmission order of the predetermined transaction nodes related to a corresponding transaction agreement. The method of claim 1,
The terminal node pool further includes a terminal management node for mapping the plurality of terminal nodes and a transaction node to which a transaction is to be transmitted from each terminal node,
The terminal management node is a dynamic blockchain system, characterized in that it performs a server function for managing not only the mapping, but also membership registration and authentication. In the method of operating a dynamic blockchain system including a terminal node pool including a plurality of terminal nodes owned by a user and a transaction node pool including a plurality of transaction nodes,
Mapping a transaction node to which a transaction is to be transmitted to each terminal node, and a transaction node to be connected next to each transaction node;
Transmitting a transaction from one of the plurality of terminal nodes to a mapped transaction node;
Asynchronously processing a transaction transmitted from the terminal node in the plurality of transaction nodes and performing virtual block consensus; And
Including; in the transaction node pool, creating a block chain by collecting at least hundreds of transactions transmitted after consensus of virtual blocks for each transaction to create a block; Including,

Asynchronously processing a transaction in the plurality of transaction nodes includes:
Mapping each transaction node to one channel,
Determining the load status of the transaction node of a specific channel based on the notification information of the transaction node transmitted each time a virtual block is agreed upon for a given transaction,
If the load of the transaction node is greater than or equal to a threshold, adding an additional transaction node to the specific channel,
Subsequently, the transaction being transmitted to the transaction node of the specific channel is dynamically distributed to the transaction node of the specific channel and the additional transaction node of the specific channel to automatically perform load balancing. How to operate a dynamic blockchain system.
delete delete

Images